This tech earnings season, Wall Streeters have been wondering one thing above all else: when will artificial intelligence begin to generate real money for anyone?
Tech giants have claimed that AI is going to disrupt every industry in the eighteen months since ChatGPT started an arms race. They have exploited this promise as rationale to spend tens of billions of dollars on data centers and chips that are required to operate large AI models. The products they’ve launched thus far, in contrast to that aim, seem very insignificant: chatbots that lack a clear way to be profitable, AI-powered customer support and coding techniques, and occasionally fictitious AI-enabled search.
However, investors are beginning to lose patience with Big Tech as they still have little to show for all the billions they have invested in terms of noteworthy revenue gains from AI or lucrative new products.
Concerns that Amazon (AMZN) is investing a lot of money on AI at a time when its core business is also facing challenges could be the main reason for the company’s less-than-impressive profits and outlook on Thursday. Due to that, the stock fell by almost 9% on Friday. The news on Thursday night that Intel is trying to rein things in by slashing $10 billion in costs and laying off tens of thousands of staff after making significant investments to adapt to the AI wave caused the company’s stock to plummet 25% on Friday.
Fears among investors can be summed up as follows: is this thing really worth anything? Is it merely an additional bright object that the industry is pursuing to revive its hopes of unending expansion before giving up and moving on to the next big thing?
During Microsoft’s earnings call, Morgan Stanley analyst Keith Weiss stated as follows: “At the moment, there’s an industry debate raging around the (capital expenditure) requirements around generative AI and whether the monetization will actually match with that.”
How long until artificial intelligence (AI) “helps revenue generation… (and) creates greater value over time, versus just cutting costs?” was the question posed to Google CEO Sundar Pichai by UBS analyst Steven Ju.
A Goldman Sachs analysis published this week questioned whether there was “too much spend, too little benefit” on generative AI.
Following their earnings announcements, shares of Google and Microsoft fell, indicating investors’ dissatisfaction that their substantial investments in artificial intelligence had not produced better-than-expected outcomes. After suffering similar displeasure from shareholders the previous quarter, Meta managed to avoid the same fate this time around by demonstrating how its AI investments were at least benefiting its main business, which included making it simple for businesses to use its AI tools to create eye-catching advertisements.
A D.A. Davidson analyst named Gil Luria told that some investors had even predicted that this would be the quarter that the big tech companies would begin to pull off their expenditures in AI infrastructure because “AI is not delivering the returns that they were expecting.”
Conversely, Google, Microsoft, and Meta all indicated that they intend to invest even more in laying the foundation for what they believe will be an AI future. The low end of the forecast was raised by $2 billion, according to Meta, which now projects full-year capital expenditures to be between $37 and $40 billion. Microsoft stated that it anticipates spending more in the 2025 fiscal year than it did in the 2024 fiscal year ($56 billion). This year, Google predicted that it will spend “at or above” $12 billion on capital expenditures every quarter. (Even for the wealthiest corporations, those are significant figures; for example, Google’s capital expenditures for the second quarter represented almost 17% of overall sales).
Additionally, tech executives have stated that they require a significant amount of extra time.
During Microsoft’s earnings call, CFO Amy Hood stated that the company anticipates that its data center investments would help monetize its artificial intelligence technologies “over the next 15 years and beyond.”
According to CFO Susan Li, Meta expects “returns from generative AI to come in over a longer period of time,” as did analysts. She continued, saying, “Gen AI is where we’re considerably sooner. Our gen AI products are not anticipated to contribute significantly to revenue in 2024. Nevertheless, we do anticipate that they will eventually present us with new revenue streams that will allow us to profit handsomely from our investment.
Many investors are uneasy with that time horizon since they are used to consistent, quarterly profit and sales growth from Silicon Valley.
A venture investment is different from an investment in a publicly traded company if the money is invested now and will be repaid in ten to fifteen years, according to Luria. We anticipate a return on investment in substantially shorter time periods for publicly traded enterprises. We aren’t currently seeing the kinds of applications or application income that would be necessary to justify making anywhere close to these investments, so that is unsettling.
Furthermore, some investors doubt that their AI investments will ever be profitable. According to Goldman Sachs analyst Jim Covello’s argument in last week’s report, “the technology isn’t designed to solve the complex problems that would justify the costs.”
Take Tesla’s AI-based “full self-driving” technology as an illustration of how long it can take for AI goods to become a reality. Since 2015, Tesla has portrayed its driver-assist technology as a crucial component of its business strategy and has made repeated assurances that it will be fully functional in a short amount of time. Even though FSD was initially made available to Tesla customers almost four years ago, it still needs a human driver who can take over in the event that something goes wrong and is frequently fraught with safety issues.
As Google CEO Sundar Pichai stated in last week’s earnings call (and as Meta CEO Mark Zuckerberg repeated in his company’s call), tech CEOs seem to be in agreement for the time being that “the risk of underinvesting is dramatically greater than the risk of overinvesting.” Construction of data centers takes time, and no business wants to lose out on the opportunity to win the AI race because they lacked the processing power. Moreover, investors will tolerate the expenditures for the time being because they are making enough money from their main operations.
However, investors will put enough pressure on tech executives to back off infrastructure spending and allow revenue growth to catch up at some point soon; Luria believes this will happen either later this year or early next.
The current strategy is, “we all have to signal that we’re willing to invest as much as we need because we want to keep this leadership position,” but eventually, Luria predicted, one of them will become so burdened with the investment that they will decide to scale back their efforts for the upcoming quarter. “Overall, this amount of investment cannot be sustained.”
